Ask AI Markdown Context and File Selection

From the FAQ:

Ask AI supports multiple AI providers, including OpenAI, Anthropic, and Google Gemini. Your Markdown content and API definitions are sent to the selected provider to generate answers. No provider uses this data for model training.

How does Ask AI in ReadMe determine which Markdown files or pages to send to the selected AI provider when a user asks a question?

Does Ask AI send all of the Markdown content in the ReadMe project, or does it identify and send only the specific files or pages that are relevant to the user’s question?

It would also be helpful to understand how Ask AI determines which Markdown pages are relevant. For example, whether it uses search, indexing, metadata, page hierarchy, or another retrieval mechanism to select the content sent to the AI provider to generate the response.

Thanks!